Proc. Natl. Acad. Sci. USA Vol. 90, pp. 3462-3465, April 1993 Neurobiology Nitric oxide/cGMP pathway stimulates phosphorylation of DARPP-32, a dopamine- and cAMP-regulated phosphoprotein, in the substantia nigra (protein phosphorylation) KANG Tsou, GRETCHEN L. SNYDER, AND PAUL GREENGARD* Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University, New York, NY 10021 Contributed by Paul Greengard, December 31, 1992

ABSTRACT DARPP-32, a dopamine- and cAMP- lular domain (7), was recently shown to produce a prominent regulated phosphoprotein of Mr 32,000, has been shown to be elevation in cGMP level and DARPP-32 phosphorylation in phosphorylated on threonine-34, both in vitro with high effi- the choroid plexus (8). The striatum and the substantia nigra ciency by cAMP-dependent and cGMP-dependent protein ki- are devoid of atrial natriuretic peptide receptors (9, 10). nases and in vivo by dopamine acting through cAMP-dependent However, the medium-sized spiny neurons of striatum and protein kinase. In the present study, we investigated the nitric their terminals in the substantia nigra contain high concen- oxide (NO)/cGMP pathway for its ability to regulate the state trations of soluble guanylyl cyclase (11), cGMP (11, 12), and of phosphorylation of DARPP-32 in slices of rat substantia cGMP-dependent protein kinase (13). The recent discovery nigra. DARPP-32 was phosphorylated on threonine-34 in these of nitric oxide (NO) as an intercellular messenger that acti- slices by sodium nitroprusside (SNP), an NO donor. The effect vates soluble guanylyl cyclase (for review, see refs. 14-17) of SNP was abolished by preincubation of the slices with and hence cGMP-dependent protein kinase (18), together hemoglobin, indicating that the effect ofSNP was due to released with the histochemical demonstration of NO synthase- NO. The same concentration ofSNP produced a 4-fold elevation containing nerve fibers in the substantia nigra (19), prompted ofthe cGMP level but did not alter the level ofcAMP. The effect us to examine the NO/cGMP pathway for a possible role in of SNP on DARPP-32 phosphorylation was mimicked by low the regulation of DARPP-32 phosphorylation. concentrations of 8-bromo-cGMP and 8-(4-chlorophenylthio)- guanosine 3',5'-cyclic monophosphate, activators of cGMP- dependent protein kinase, but not by low concentrations of MATERIALS AND METHODS 8-bromo-cAMP, an activator of cAMP-dependent protein ki- Materials. Sodium nitroprusside (SNP), L-arginine, nase. The data indicate a physiological role for the NO/cGMP 8-bromo-cAMP, 8-bromo-cGMP, RPMI 1640 balanced salt pathway in the regulation of DARPP-32 phosphorylation in solution (RPMI medium), bovine Hb, and sodium dithionite nerve terminals of striatonigral neurons. The results provide were obtained from Sigma. 8-(4-Chlorophenylthio)guanosine further evidence that the state of phosphorylation ofDARPP-32 3',5'-cyclic monophosphate (8-pCPT-cGMP) was purchased represents an important mechanism for integration of informa- from BioLog Life Science Institute (Bremen, Germany), tion arriving at striatonigral neurons via a variety of neuronal bicinchoninic acid reagent and goat anti-mouse horseradish pathways. peroxidase-linked antibody were from Pierce, Bradford pro- tein assay reagent was obtained from Bio-Rad, nitrocellulose DARPP-32, a dopamine- and cAMP-regulated phosphopro- filters (pore size, 0.2 am) were from Schleicher & Schuell, tein of Mr 32,000 on SDS/polyacrylamide gels, is a cytosolic forskolin was purchased from Calbiochem, and N-methyl-D- phosphoprotein that is highly enriched in the medium-sized aspartic acid (NMDA) was from Tocris (Essex, England). spiny neurons of the striatum and in the striatonigral nerve cAMP and cGMP RIA kits and enhanced chemiluminescence terminals in the substantia nigra (1-3). Previous studies have detection kits were obtained from Amersham. shown that dopamine, by increasing levels of cAMP and Hb was prepared by adding a 10-fold molar excess of activating cAMP-dependent protein kinase, stimulates the sodium dithionite to a 1 mM solution ofcommercial Hb, 75% phosphorylation of DARPP-32 in these brain regions (4). of which is methemoglobin. Sodium dithionite was removed Phosphorylation of a single threonine residue, in position 34, by dialysis against nanopure distilled water. The freshly by cAMP-dependent protein kinase, converts DARPP-32 Hb into a potent inhibitor of protein phosphatase 1 (5). Kinetic prepared solution was divided into aliquots, frozen at studies of purified DARPP-32 indicate that this threonine -20°C, and stored for up to 14 days (20). residue is also an excellent substrate for phosphorylation by Assay of Phospho-DARPP-32. A phosphorylation state- cGMP-dependent protein kinase in vitro (6). Those results specific monoclonal antibody was raised against a DARPP-32 raised the possibility that the physiological actions of peptide containing phosphothreonine-34, the site phosphor- DARPP-32 may be regulated through its phosphorylation by ylated by cAMP-dependent and cGMP-dependent protein either cAMP-dependent protein kinase or cGMP-dependent kinases. This antibody, which has been characterized previ- protein kinase. However, no signal transduction pathway in ously (8), was used to assay the amount of DARPP-32 the brain has yet been demonstrated to regulate the state of phosphorylated at threonine-34. phosphorylation of DARPP-32 through the activation of Preparation of Nigral Slices. Female Sprague-Dawley rats cGMP-dependent protein kinase. (180-200 g) were stunned and decapitated. The brain was Atrial natriuretic peptide, a hormone that activates a rapidly removed and immersed in ice-cold RPMI medium. membrane-bound receptor with a guanylyl cyclase intracel- Abbreviations: DARPP-32, dopamine- and cAMP-regulated phos- phoprotein of Mr 32,000; SNP, sodium nitroprusside; 8-pCPT- The publication costs of this article were defrayed in part by page charge cGMP, 8-(4-chlorophenylthio)guanosine 3',5'-cyclic monophos- payment. This article must therefore be hereby marked "advertisement" phate; NMDA, N-methyl-D-aspartic acid. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed.

3462 Downloaded by guest on September 26, 2021 Neurobiology: Tsou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 3463 The midbrain was isolated by cutting the brainstem coronally Table 1. Effect of test substances on phosphorylation of between the mammillary body and the pons. The midbrain DARPP-32 in slices of substantia nigra was laid on a flat glass surface on ice, and the substantia nigra DARPP-32 was dissected out and further divided into five or six small phosphorylation, slices. The nigral slices from each side were placed into Test substance % control separate tubes. The slices were preincubated in warm (30°C), oxygenated RPMI medium. After a 45-min preincubation, the SNP (1 mM) 269 ± 38 (13) solution was replaced with fresh RPMI medium containing Hb (20,uM) 104 ± 13 (5) test substances, and incubation was carried out for 10 min. At SNP (1 mM) plus Hb (20 ,uM) 105 ± 14 (5) the was removed, and the 8-Br-cGMP (1 mM) 658 ± 55 (5) the end of the incubation, medium 8-Br-cGMP (0.1 mM) 409 ± 87 (8) tissues were rapidly frozen in liquid nitrogen and stored at 8-pCPT-cGMP (1 mM) 703 ± 91 (4) -70°C until assayed. 8-pCPT-cGMP (0.1 mM) 361 ± 63 (4) Immunoblotting. Nigral slices were homogenized by son- 8-Br-cAMP (1 mM) 784 ± 64 (4) ication in boiling 1% (wt/vol) SDS. The protein concentration 8-Br-cAMP (0.1 mM) 200 ± 38 (5) was determined with a bicinchoninic acid-based assay (21), using bovine serum albumin as a standard. Aliquots of the Experimental procedures were as described in Materials and homogenates containing equal amounts of protein (180-250 Methods and in the legend to Fig. 1. Data represent mean ± SEM for ,ug) were removed, and 0.25 vol of five times concentrated the number of experiments indicated in parentheses. stop solution was added [final concentrations: 3% SDS, 62 buffer instead of in boiling SDS. An aliquot was taken for mM Tris-HCl (pH 6.8), 0.3 M 2-mercaptoethanol, 5% (vol/ protein concentration measurement using the Bradford vol) glycerol, and traces ofbromophenol blue]. Proteins were method (24). The remainder of the solution was then heated separated by SDS/15% PAGE (22). After electrophoresis, for 3 min in a boiling water bath to denature the protein. After proteins in the gel were transferred to nitrocellulose filters centrifugation to precipitate protein, the supematant was (23). Immunoblotting was performed at room temperature divided and used for cGMP and cAMP assays according to using 4% (wt/vol) nonfat dry milk as blocking agent. After the Amersham protocols. blocking for 2 hr, the nitrocellulose filters were incubated with the monoclonal antibody specific for phospho- DARPP-32 (1:100) for 2 hr, washed, and incubated with a goat RESULTS anti-mouse horseradish peroxidase antibody (1:3000) for 30 Phosphorylation of DARPP-32. SNP in- min. The enhanced chemiluminescence Western blotting Effect of SNP on detection-method was used to reveal antibody binding. The creased phosphorylation of DARPP-32 on threonine-34 in filters were dried, wrapped in Saran Wrap, and exposed to slices of substantia nigra (Fig. 1). The level of phospho- x-ray film. DARPP-32 was 2- to 3-fold higher in SNP-treated slices as The phospho-DARPP-32 bands on the x-ray films were compared with control slices (Table 1). In the same series of quantitated by using a Bio-Rad model 620 video densitometer experiments in which SNP was examined for a possible effect and Bio-Rad one-dimensional MOLECULAR ANALYST soft- on DARPP-32 phosphorylation, forskolin, which increases ware. Results were calculated as a percentage of the control. intracellular cAMP and activates cAMP-dependent protein cGMP and cAMP RIA. Preparation and treatment of nigral kinase, also increased DARPP-32 phosphorylation, in agree- slices was the same as for the phosphorylation experiments, ment with previous results (25). 8-Br-cGMP, which activates except that the slices were sonicated in 4 mM Tris/EDTA cGMP-dependent protein kinase, also increased DARPP-32 phosphorylation (Fig. 1; see below). a. The effect of various concentrations of SNP on DARPP-32 0 0 phosphorylation is shown in Fig. 2. A maximal effect of SNP was observed at a final concentration of 1 mM. The effect of L- -e I -6- L cV C .co 0C incubating slices of substantia nigra with SNP is shown as a o o z 0 function of incubation time in Fig. 3. The effect of SNP was ) tl co C' maximal at an incubation time of 10 min and declined x 106 thereafter. I 80 C 300 0 I-o 1- 49 0

--DARPP-32 N -jCC 32 cv0C) 200 27 - .'::::. -J a 0 .b 4 - I.pmr a0

Q o FIG. 1. Effects offorskolin (10 ,uM), 8-Br-cGMP (4 mM), and SNP 0 (1 mM) on DARPP-32 phosphorylation in slices ofrat substantia nigra. Slices were preincubated in oxygenated RPMI medium at 30°C for 45 100 min. The medium was replaced by fresh solution with or without the 0 0.5 1 2 3 indicated test substance, and incubation was carried for an additional SNP (mM) 10 min. The tissues were frozen in liquid nitrogen and stored at -70°C until assayed. Proteins were separated by SDS/PAGE and transferred FIG. 2. Effect of various concentrations of SNP on DARPP-32 to a nitrocellulose filter. Phosphorylated DARPP-32 was detected phosphorylation in slices of substantia nigra. Experimental proce- using a phosphorylation state-specific antibody. Other experimental dures were as described in Materials and Methods and in the legend details were as described in Materials and Methods. to Fig. 1. Downloaded by guest on September 26, 2021 3464 Neurobiology: Tsou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 300 0 1*20r cGMP 36 cAMP 1- I 0 1.00[ 30 atcs 0~ 200 24 0) 0.80[ cm0~ E '- 0.60 18 T E r M0. :a 100 ' 0.40 T 12 m 0.20[ 6 0.Go 0. 0 0.00L 0 0 10 20 30 controI SNP controi SNP FIG. 5. Effect of SNP on cGMP and cAMP levels in slices of Time (min) substantia nigra. Cyclic nucleotide levels were measured by RIA after incubation of slices in the presence of SNP (1 mM) for 10 min. FIG. 3. Effect of SNP on DARPP-32 phosphorylation in slices of (Difference between control and SNP-treated slices by Student's t substantia nigra, as a function of incubation time. Experimental test: cGMP, n = 6, P < 0.01; cAMP, n = 4, P > 0.05.) Experimental procedures were as described in Materials and Methods and in the procedures were as described in Materials and Methods and in the legend to Fig. 1. legend to Fig. 1.

Effect of Hb Pretreatment on SNP-Stimulated Phosphory- Effect of Cyclic Nucleotide Derivatives on Phosphorylation of lation of DARPP-32. To verify that the SNP effect was due to DARPP-32. Evidence in support of the idea that SNP in- NO released into the incubation medium, Hb was added to creased the phosphorylation of DARPP-32 by virtue of acti- the medium in the absence or presence of SNP. Preincuba- vating cGMP-dependent protein kinase was obtained in ex- tion of slices with 20 u.M Hb for 10 min prior to addition of periments utilizing 8-Br-cGMP and 8-pCPT-cGMP, two stable SNP abolished the SNP-stimulated phosphorylation of analogs of cGMP that selectively activate cGMP-dependent DARPP-32 (Fig. 4 and Table 1). In the absence of SNP, Hb protein kinase (18, 26). At a concentration of 1 mM, each of had no effect on the level of phosphorylation of DARPP-32. these compounds markedly stimulated DARPP-32 phosphor- Effect of SNP on cGMP and cAMP Levels. One of the major ylation, as did 8-Br-cAMP, an analog ofcAMP that selectively biological effects of NO in a variety of tissues (17) is to activates cAMP-dependent protein kinase (Fig. 6 and Table 1). stimulate soluble guanylyl cyclase, thereby increasing the In contrast, at a concentration of 0.1 mM, each of the cGMP formation of cGMP. As shown in Fig. 5, incubation of slices analogs caused a substantially greater increase in DARPP-32 of substantia nigra with 1 mM SNP for 10 min increased the phosphorylation than did 8-Br-cAMP. cGMP level -4-fold, from 0.28 + 0.10 to 1.01 ± 0.15 pmol/mg of protein (mean ± SEM, n = 6; P < 0.01). These DISCUSSION results are consistent with the results of an immunocyto- Recent studies indicate that the NO/cGMP pathway is an that SNP treatment chemical study showing greatly increased important intercellular signaling mechanism in a wide variety the content in nerve cGMP fibers of the substantia nigra (J. of tissues in the periphery as well as in the central nervous de Vente, personal communication). The basal level ofcAMP system (14-17). Protein phosphorylation is one plausible in the nigral slices was much higher (22.2 ± 7.4 pmol/mg of mechanism through which activation of soluble guanylyl protein; mean ± SEM, n = 4) than that of cGMP. SNP cyclase by NO could regulate target cells. The present study treatment had no significant effect on the level of cAMP (Fig. demonstrated that SNP increased the phosphorylation of 5). DARPP-32 on threonine-34 in tissue slices prepared from rat substantia nigra. To our knowledge, these data provide the first demonstration of a NO-mediated effect on protein phos- phorylation in the central nervous system. The SNP-induced O LI) increase in DARPP-32 phosphorylation in nigral slices was abolished by pretreatment of the slices with Hb, a compound 0 Z +mv0 z D m I 1 3 4 5 6 7 m co x I do 2 b co 106 __ -,-- _ - 49 ;o x I, 80 x~~~~~~~~~~"! 8 0ffimk'~1ik ...... " IMP

32 - - DARPP-32 49

0

32 u- DARPP-32 FIG. 4. Effect of preincubation with Hb on SNP-stimulated phosphorylation of DARPP-32 in slices of substantia nigra. Slices 1 mM 0.1 mM were preincubated in oxygenated RPMI medium at 30°C for 30 min. Hb was then added and preincubation was continued for an addi- FIG. 6. Comparison of the effects of cGMP and cAMP deriva- tional 10 min. After this 40-min preincubation period, SNP was added tives on phosphorylation of DARPP-32 in slices of substantia nigra. and incubation was carried out for 10 min. The test substances were 8-Br-cGMP (lanes 2 and 5), 8-pCPT-cGMP (lanes 3 and 6), and present as indicated, at the following final concentrations: SNP, 1 8-Br-cAMP (lanes 4 and 7) were present in final concentrations of 1 mM; Hb, 20 ,uM; 8-Br-cGMP, 100 ,M. Other experimental proce- mM or 0.1 mM as indicated. Lane 1, control. Other experimental dures were as described in Materials and Methods and in the legend procedures were as described in Materials and Methods or in the to Fig. 1. legend to Fig. 1. Downloaded by guest on September 26, 2021 Neurobiology: Tsou et al. Proc. Natl. Acad. Sci. USA 90 (1993) 3465 that, by complexing with NO, prevents the activation of American Parkinson's Disease Association and funding from Elf- guanylyl cyclase by NO. The hypothesis that NO regulates Aquitaine, Inc. (Montpellier, France). phosphorylation of DARPP-32 by increasing the activity of cGMP-dependent protein kinase is strengthened by the ob- 1. Walaas, S. I. & Greengard, P. (1984) J. Neurosci. 4, 84-98. servation that SNP treatment increased cGMP levels. The 2. Ouimet, C. C., Miller, P. E., Hemmings, H. C., Jr., Walaas, S. I. & Greengard, P. (1984) J. Neurosci. 4, 111-124. possible involvement of cGMP-dependent protein kinase in 3. Hemmings, H. C., Jr., & Greengard, P. (1986) J. Neurosci. 6, the SNP-stimulated phosphorylation of DARPP-32 is also 1469-1481. supported by experiments with the cGMP analogs 8-Br- 4. Walaas, S. I., Aswad, D. W. & Greengard, P. (1983) Nature cGMP and 8-pCPT-cGMP, showing that activation ofcGMP- (London) 301, 69-71. dependent protein kinase alone is sufficient to increase the 5. Hemmings, H. C., Jr., Greengard, P., Tung, H. Y. L. & Co- phosphorylation state of DARPP-32. hen, P. (1984) Nature (London) 310, 503-505. The present results provide evidence that the NO/cGMP 6. Hemmings, H. C., Jr., Nairn, A. C. & Greengard, P. (1984) J. pathway regulates the phosphorylation of DARPP-32 in Biol. Chem. 259, 14491-14497. 7. Maack, T. (1992) Annu. Rev. Physiol. 54, 11-27. nerve terminals of striatonigral neurons. However, the iden- 8. Snyder, G. L., Girault, J.-A., Chen, J. Y. C., Czernik, A. J., tity of the cell type(s) that participates in endogenous for- Kebabian, J. W., Nathanson, J. A. & Greengard, P. (1992) J. mation of NO is unclear. Many nerve fibers in the substantia Neurosci. 12, 3071-3083. nigra express NO synthase as shown by NADPH diaphorase 9. Quirion, R., Dalpe, M. & Dam, T.-V. (1986) Proc. Natl. Acad. histochemistry (19, 27, 28). Cholinergic fibers innervating the Sci. USA 83, 174-178. substantia nigra and originating in the laterodorsal pontine 10. Mantyh, C. R., Kruge, R. L., Breche, N. C. & Mantyh, P. W. tegmentum (28-31) have been shown to coexpress NO syn- (1987) Brain Res. 412, 329-342. 11. Ariano, M. A. (1983) Neuroscience 10, 707-723. thase (28) and are a plausible candidate for the site of NO 12. Ariano, M. A. & Ufkes, S. K. (1983) Neuroscience 9, 23-29. formation. 13. Walaas, S. I., Girault, J.-A. & Greengard, P. (1989) J. Mol. The nature ofthe transmitter(s) that activates NO synthase Neurosci. 1, 243-250. in nerve fibers in the substantia nigra also remains to be 14. Moncada, S., Palmer, R. M. J. & Higgs, E. A. (1991) Phar- clarified. Previous studies using slices of rat cerebellum, a macol. Rev. 43, 109-142. brain region with many NO synthase-containing neurones 15. Garthwaite, J. (1991) Trends NeuroSci. 11, 60-67. and fibers (16, 19, 32), showed that L-arginine, the NO 16. Bredt, D. S. & Snyder, S. H. (1992) Neuron 8, 3-11. 17. Lowenstein, C. J. & Snyder, S. H. (1992) Cell 70, 705-707. precursor (33), did not elevate cGMP levels when treated 18. Geiger, J., Nolte, C., Butt, E., Sage, S. 0. & Walter, U. (1992) alone but did potentiate the cGMP-elevating effect of the Proc. Natl. Acad. Sci. USA 89, 1031-1035. glutamatergic agonist NMDA (34). In the present study, no 19. Vincent, S. R. & Kimura, H. (1992) Neuroscience 46, 755-784. effect of L-arginine alone or L-arginine plus NMDA on 20. Martin, W., Villani, G. M., Jothianandan, D. & Furchgott, R. DARPP-32 phosphorylation was observed in slices of sub- (1985) J. Pharmacol. Exp. Ther. 232, 708-716. stantia nigra. These results with the substantia nigra suggest 21. Smith, P. K., Krohn, R. I., Hermanson, C. T., Mallia, A. K., that NMDA-receptive neurons may not be involved or that Gartner, F. H., Provenzans, M. S., Fugimoto, E. K., Goeke, N. M., Olson, B. J. & Klenk, D. C. (1985) Anal. Biochem. 150, the amount of NO synthesis was below the threshold for the 76-85. activation of the cGMP/cGMP-dependent protein kinase 22. Laemmli, U. K. (1970) Nature (London) 227, 680-685. pathway under the experimental conditions tested. 23. Renart, J. & Sandoval, I. V. (1984) Methods Enzymol. 104, The results of the present study support the concept that 455-460. DARPP-32 represents a final common pathway through 24. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. which many nerve cells and neurotransmitters exert their 25. Halpain, S., Girault, J.-A. & Greengard, P. (1990) Nature actions on medium-sized spiny neurons. For instance, dopa- (London) 343, 369-372. mine, acting on D1 receptors, elevates intracellular cAMP, 26. Butt, E., Nolte, C., Schulz, S., Beltman, J., Beavo, J. A., to activation of kinase and Jastorff, B. & Walter, U. (1992) Biochem. Pharmacol. 43, leading cAMP-dependent protein 2591-2600. phosphorylation of DARPP-32 (4). Glutamate, acting on 27. Hope, B. T., Michael, G. J., Knigge, K. M. & Vincent, S. R. NMDA receptors, elevates intracellular calcium, leading to (1991) Proc. Natl. Acad. Sci. USA 88, 2811-2814. activation of calcium/calmodulin-dependent phosphatase 28. Vincent, S. R., Satoh, K., Armstrong, D. M., Panula, P., Vale, and dephosphorylation of DARPP-32 (25). y-Aminobutyric W. & Fibiger, H. C. (1986) Neuroscience 17, 167-182. acid, acting on type A y-aminobutyric acid receptors, phos- 29. Moon Edley, S. & Graybiel, A. M. (1983) J. Comp. Neurol. phorylates DARPP-32 through an unknown intracellular sig- 217, 187-215. nal transduction mechanism (35). The present results indicate 30. Sugimoto, T. & Hattori, T. (1984) Neuroscience 11, 931-946. that NO, an intercellular messenger, elevates intracellular 31. Satoh, K. & Fibiger, H. C. (1986) J. Comp. Neurol. 253, cGMP, leading to activation of cGMP-dependent protein 277-302. 32. Bredt, D. S., Hwang, P. M. & Snyder, S. H. (1990) Nature kinase and phosphorylation of DARPP-32. The state of (London) 347, 768-770. phosphorylation of DARPP-32 appears to provide an impor- 33. Knowles, R. G., Palacios, M., Palmer, R. M. J. & Moncada, S. tant mechanism for integration of information converging on (1989) Proc. Natl. Acad. Sci. USA 86, 5159-5162. striatonigral neurons from a variety of neuronal pathways. 34. Garthwaite, J., Garthwaite, G., Palmer, R. M. J. & Moncada, S. (1989) Eur. J. Pharmacol. 172, 413-416. This work was supported by U.S. Public Health Service Grant 35. Snyder, G. L., Fisone, G. & Greengard, P. (1992) Soc. Neu- MH-40899. G.L.S. was the recipient of a research grant from the rosci. Abstr. 18, 104s. Downloaded by guest on September 26, 2021